Search

Focal traumatic injury to the cerebral cortex is associated with early activation of the neuronal isoform of nitric oxide synthase (nNOS), where high concentrations of nitric oxide-derived free radicals elicit extensive DNA damage. Subsequent activation of the nuclear repair enzyme poly(ADP-ribose) polymerase (PARP) causes a severe energy deficit leading to the ultimate demise of affected neurons. Little is known about the temporal relationship of nNOS and PARP activation and the neuroprotective efficacy of their selective blockade in traumatic brain injury. To determine the relationship of nNOS and PARP activation, brain injury was induced by cryogenic lesion to the somatosensory cortex applying a pre-cooled cylinder after trephination for 6 s to the intact dura mater. Pre-treatment with 3-bromo-7-nitroindazole (BrNI; 25 mg/kg, i.p.), and pre- or combined pre- and post-treatment with 3-aminobenzamide (AB; 10 mg/kg (i.c.v.) or 10 mg/kg/h (i.p.)) were used to inhibit nNOS and PARP, respectively. Cold lesion-induced changes in the somatosensory cortex and neuroprotection by BrNI and AB were determined using immunocytochemistry and immunodot-blot for detection of poly(ADP-ribose; PAR), the end-product of PARP activation, and the triphenyltetrazolium-chloride assay to assess lesion volume. PAR immunoreactivity reached its peak 30 min post-lesion and was followed by gradual reduction of PAR immunolabeling. BrNI pre-treatment significantly decreased the lesion-induced PAR concentration in damaged cerebral cortex. Pre-treatment by i.c.v. infusion of AB markedly diminished cortical PAR immunoreactivity and significantly reduced the lesion volume 24 h post-injury. In contrast, i.p. AB treatment remained largely ineffective.In conclusion, our data indicate early activation of PARP after cold lesion that is, at least in part, related to nNOS induction and supports the relevance of nNOS and/or PARP inhibition to therapeutic approaches of traumatic brain injury. [Copyright &y& Elsevier]

Biodegradable collagen scaffolds are used clinically for oral soft tissue augmentation to support wound healing. This study sought to provide a novel ex vivo model for analyzing healing kinetics and gene expression of primary human gingival fibroblasts (hGF) within collagen scaffolds. Sponge type and gel type scaffolds with and without platelet-derived growth factor-BB (PDGF) were assessed in an hGF containing matrix. Morphology was evaluated with scanning electron microscopy, and hGF metabolic activity using MTT. We quantitated the population kinetics within the scaffolds based on cell density and distance from the scaffold border of DiI-labled hGFs over a two-week observation period. Gene expression was evaluated with gene array and qPCR. The sponge type scaffolds showed a porous morphology. Absolute cell number and distance was higher in sponge type scaffolds when compared to gel type scaffolds, in particular during the first week of observation. PDGF incorporated scaffolds increased cell numbers, distance, and formazan formation in the MTT assay. Gene expression dynamics revealed the induction of key genes associated with the generation of oral tissue. DKK1, CYR61, CTGF, TGFBR1 levels were increased and integrin ITGA2 levels were decreased in the sponge type scaffolds compared to the gel type scaffold. The results suggest that this novel model of oral wound healing provides insights into population kinetics and gene expression dynamics of biodegradable scaffolds.

Objectives: The aim was to test, whether or not soft tissue volume augmentation with a specifically designed collagen matrix (CM), leads to ridge width gain in chronic ridge defects similar to those obtained by an autogenous subepithelial connective tissue graft (SCTG)., Material and Methods: In six dogs, soft tissue volume augmentation was performed by randomly allocating three treatment modalities to chronic ridge defects [CM, SCTG and sham-operated control (Control)]. Dogs were sacrificed at 28 (n = 3) and 84 days (n = 3). Descriptive histology and histomorphometric measurements were performed on non-decalcified sections., Results: SCTG and CM demonstrated favourable tissue integration, and subsequent re-modelling over 84 days. The overall mean amount of newly formed soft tissue (NMT) plus bone (NB) amounted to 3.8 ± 1.2 mm (Control), 6.4 ± 0.9 mm (CM) and 7.2 ± 1.2 mm (SCTG) at 28 days. At 84 days, the mean NMT plus NB reached 2.4 ± 0.9 mm (Control), 5.6 ± 1.5 mm (CM) and 6.0 ± 2.1 mm (SCTG). Statistically significant differences were observed between CM/SCTG and Control at both time-points (p < 0.05)., Conclusion: Within the limits of this animal model, the CM performed similar to the SCTG, based on histomorphometric outcomes combining NB and NMT., (© 2011 John Wiley & Sons A/S.)

This study evaluates a new collagen matrix that is designed with a bilayered structure in order to promote guided tissue regeneration and integration within the host tissue. This material induced a mild tissue reaction when assessed in a murine model and was well integrated within the host tissue, persisting in the implantation bed throughout the in vivo study. A more porous layer was rapidly infiltrated by host mesenchymal cells, while a layer designed to be a barrier allowed cell attachment and host tissue integration, but at the same time remained impermeable to invading cells for the first 30 days of the study. The tissue reaction was favorable, and unlike a typical foreign body response, did not include the presence of multinucleated giant cells, lymphocytes, or granulation tissue. In the context of translation, we show preliminary results from the clinical use of this biomaterial applied to soft tissue regeneration in the treatment of gingival tissue recession and exposed roots of human teeth. Such a condition would greatly benefit from guided tissue regeneration strategies. Our findings demonstrate that this material successfully promoted the ingrowth of gingival tissue and reversed gingival tissue recession. Of particular importance is the fact that the histological evidence from these human studies corroborates our findings in the murine model, with the barrier layer preventing unspecific tissue ingrowth, as the scaffold becomes infiltrated by mesenchymal cells from adjacent tissue into the porous layer. Also in the clinical situation no multinucleated giant cells, no granulation tissue and no evidence of a marked inflammatory response were observed. In conclusion, this bilayered matrix elicits a favorable tissue reaction, demonstrates potential as a barrier for preferential tissue ingrowth, and achieves a desirable therapeutic result when applied in humans for soft tissue regeneration.

Gingival cells of the oral connective tissue are exposed to complex mechanical forces during mastication, speech, tooth movement and orthodontic treatments. Especially during wound healing following surgical procedures, internal and external forces may occur, creating pressure upon the newly formed tissue. This clinical situation has to be considered when developing biomaterials to augment soft tissue in the oral cavity. In order to pre-evaluate a collagen sponge intended to serve as a substitute for autogenous connective tissue grafts (CTGs), a dynamic bioreactor system was developed. Pressure and shear forces can be applied in this bioreactor in addition to a constant medium perfusion to cell-material constructs. Three-dimensional volume changes and stiffness of the matrices were analyzed. In addition, cell responses such as cell vitality and extracellular matrix (ECM) production were investigated. The number of metabolic active cells constantly increased under fully dynamic culture conditions. The sponges remained elastic even after mechanical forces were applied for 14 days. Analysis of collagen type I and fibronectin revealed a statistically significant accumulation of these ECM molecules (P < 0.05-0.001) when compared to static cultures. An increased expression of tenascin-c, indicating tissue remodeling processes, was observed under dynamic conditions only. The results indicate that the tested in vitro cell culture system was able to mimic both the biological and mechanical environments of the clinical situation in a healing wound., (© 2010 Wiley Periodicals, Inc.)

Objectives: The aim was to test whether or not soft tissue augmentation with a newly developed collagen matrix (CM) leads to volume gain in chronic ridge defects similar to those obtained by an autogenous subepithelial connective tissue graft (SCTG)., Material and Methods: In six dogs, soft tissue volume augmentation was performed by randomly allocating three treatment modalities to chronic ridge defects (CM, SCTG, sham-operated control). Impressions were taken before augmentation (baseline), at 28, and 84 days. The obtained casts were optically scanned and the images were digitally analysed. A defined region of interest was measured in all sites and the volume differences between the time points were calculated., Results: The mean volume differences per area between baseline and 28 days amounted to a gain of 1.6 mm (CM; SD+/-0.9), 1.5 mm (SCTG; +/-0.1), and a loss of 0.003 mm (control; +/-0.3). At 84 days, the mean volume differences per area to baseline measured a gain of 1.4 mm (CM; +/-1.1), 1.4 mm (SCTG; +/-0.4), and a loss of 0.3 mm (control; +/-0.3). The differences between CM and SCTG were statistically significant compared with control at 28 and 84 days (p<0.001)., Conclusion: Within the limits of this animal study, the CM may serve as a replacement for autogenous connective tissue.

The aim of the present study was to measure the therapeutic effects of bradykinin antagonists on lesion volume and brain swelling induced by cold injury in the parietal cortex of rat and mouse, respectively. Cold lesion was induced by application of a precooled (-78 degrees C) copper cylinder (3 mm diameter) to the intact dura of rat and mouse for 6 and 30 sec, respectively. At 24 h after the injury, the brains were removed and lesion volume was determined by the triphenyltetrazolium chloride method in rats. In the mouse, brain swelling was expressed as percentage increase in weight of the injured hemisphere which is compared to the contralateral side. After a subcutaneous priming dose of 18 microg/kg, a 1-h pretreatment and 24-h posttreatment using osmotic minipumps (300 ng/kg x min) was applied. Hoe140, a bradykinin receptor 2 antagonist, revealed a 19% reduction of lesion volume (p < 0.05) in the rat and a 14% diminution of brain swelling (p < 0.05) in the mouse. In contrast, the bradykinin receptor 1 antagonist, B 9858, had no effect on lesion volume compared to sham treated rats. When B 9858 was given in combination with Hoe140, a significant reduction in lesion volume was seen which was equivalent to and not different from that seen with Hoe140 alone in the rat. We conclude that brain injury after cold lesion is partially mediated by bradykinin and can be successfully treated with B2 antagonists.

The interaction between uridine-5'-triphosphate (UTP) and prostanoids was studied in isolated rat middle cerebral arteries (MCAs). The strong contractions in MCA segments induced by UTP were weakened significantly by indomethacin and more markedly by the thromboxane receptor antagonist ICI 192605. Thromboxane A(2) (TXA(2)) release by MCAs was below the detection limit of the chemiluminescence enzyme immunoassay, but increased TXA(2) formation was detected in basilar arteries in the presence of UTP. Prostacyclin (PGI(2)) formation by MCAs also increased in the presence of UTP. These results suggest that UTP stimulates the release of both TXA(2) and PGI(2) from the rat MCA but the vascular effect of TXA(2) is dominant.

The aim of the present study was to investigate whether endothelin-1 (ET-1) in cerebral arteries is inhibited by the new, non-peptidergic ET(A) receptor antagonist Ro 61-1790 and, if it is, whether that inhibition reduces the lesion volume induced by cold injury in the parietal cortex. In vitro experiments were performed by measuring the isometric contractions of the rat middle cerebral and basilar arteries. A cold lesion was induced in vivo by the application of a pre-cooled (-78 degrees C) copper cylinder (diameter 3 mm) to the intact dura of rats for 6 s. After 24 h, lesion volume was determined by the triphenyltetrazolium method. In vitro, ET-1 (10(-12) - 3x10(-7) M) caused a dose-dependent contraction under resting conditions in the middle cerebral and basilar arteries of control rats. Ro 61-1790 (3x10(-9) M, 10(-7) M) shifted the concentration-effect curves for ET-1 in a parallel fashion (Emax unaltered). Post-treatment with Ro 61-1790 (10(-7)-10(-5) M) also inhibited the prior contraction elicited by ET-1 (3x10(-9) M) significantly. In vitro ET-1 application 3 h after the intracerebroventricular in vivo administration of Ro 61-1790 showed that the antagonist had reached the arteries and was bound to their ET(A) receptors. Intracerebroventricular pre-treatment of Ro 61-1790 reduced significantly the lesion volume by 23% after the injury. We conclude that ET-1 is involved in the development of secondary brain damage and that intracerebroventricular treatment with Ro 61-1790 reduces the size of the brain lesion caused by cold injury.

The effects of nitric oxide synthase (NOS) blockade on the cerebrocortical microcirculation were investigated under physiological conditions in anesthetized ventilated rats using laser-Doppler (LD) flowmetry. LD flow values of the parietal cortex were determined before and after systemic administration of the NOS inhibitor N(G)-nitro-L-arginine-methyl-esther. NOS blockade reduced the LD flow significantly and the magnitude of the reduction was in close correlation with the baseline value. Synchronized sinus-wave-like LD flow oscillations were observed frequently after NOS inhibition and their appearance was also dependent on the high baseline flow values. These results indicate marked, baseline-dependent differences in the cerebrocortical blood flow response to the inhibition of the nitric oxide pathway, and may suggest that areas with high resting red blood cell flow express high NOS activity.

The function of the neuronal isoform of nitric oxide synthase (nNOS) was studied by comparing the effects of the specific nNOS blocker 7-nitro indazole monosodium salt (7-NINA) with that of the general NOS inhibitor N(G)-nitro-L-arginine (L-NA) in isolated rat basilar arteries (BAs). 7-NINA had no significant effect on the resting tone of the vessels, while both L-NA and 1H-[1,2,4]oxadiazolo[4, 3-a]quinoxalin-1-one (ODQ), a selective inhibitor of the soluble guanylyl cyclase, induced contraction. The relaxant effect of bradykinin was attenuated in the presence of L-NA but was not changed by 7-NINA. In contrast, 7-NINA markedly reduced the acetylcholine-induced, endothelium-dependent relaxation. These results demonstrate that nNOS contributes significantly to the relaxant effect of acetylcholine, indicating the functional importance of this isoenzyme in the cerebrovascular endothelium.

Purpose: The aim of the present study was to assess the effects of neuronal nitric oxide synthase (NOS I) inhibitors and a combination of NOS I and NOS II inhibitors on lesion volume after experimental brain injury., Methods: Cold lesion of the brain was induced by application of a precooled (.... 78 °C) copper cylinder to the intact dura of the rat for 6 s. Brains were removed 24 h after the injury and lesion volume determined using the triphenyltetrazolium-chloride method., Results: The specific NOS I inhibitor 3-bromo-7-nitroindazole (Br-7-NI) reduced lesion volume significantly by 21 % compared with the vehicle control. In contrast, 7-nitroindazole had no effect on lesion volume. When aminoguanidine, a specific NOS II inhibitor, was adminis-tered after Br-7-NI, lesion volume was significantly reduced but not significantly more than with either compound alone., Conclusion: Brain injury after cold lesion is partly mediated by NOS I activity and can be attenuated successfully with Br-7-NI, while coin-hibition of NOS II does not improve the outcome significantly.

We used in situ hybridization, RT PCR and immunohistochemistry to study the time course of expression and the cellular localization of inducible nitric oxide synthase (iNOS) and interleukin-1beta (IL-1beta) during the first 7 days after induction of a standardized cryogenic lesion on the right parietal cortex in male rats. Cryogenic lesion induced iNOS mRNA in the lesioned hemisphere after 6 to 72 h with a maximum (15+/-2 cells/mm2, n=4, p<0.01 vs. sham) at 24 h. Microglia, invading monocytes and granulocytes in and around the lesion expressed iNOS immunoreactivity starting at 12 h and peaking (29+/-10 cells/mm2, n=4, p<0.05 vs. sham) at 24 h after lesion. Induction of IL-1beta mRNA expression was immediate with a peak (9+/-1 cells/mm2, n=4, p<0.01 vs. sham) at 24 h after cryogenic lesion. The number of round cells with IL-1beta immunoreactivity around the lesion was maximal (8+/-2 cells/0.1 mm2, n=3, p<0.01 vs. sham) at 24 h. A weak astrocytic expression of IL-1beta-immunoreactivity was seen in sham animal brains. Astrocytic IL-1beta-expression was significantly increased in the lesion hemisphere and both hippocampi. Interleukin converting enzyme (ICE) was expressed in astrocytes and microglia around the lesion 6 h after injury. The number of ICE immunoreactive cells (8+/-2 cells/0. 1 mm2, n=3, p<0.05 vs. sham) peaked at 72 h after lesion. Neuronal expression of ICE and IL-1beta was seen in the lesion periphery 72 h and 7 days after injury. At this time, morphological features of apoptosis were evident in cells in the lesion periphery. The data indicate an early activation of microglia and monocyte invasion into the lesion hemisphere leading to multicellular expression of iNOS, ICE, and IL-1beta. These events may contribute to the expansion of neuronal damage after brain injury., (Copyright 1999 Elsevier Science B.V.)

The role of the neuronal isoform of the nitric oxide (NO) synthase (nNOS) in the regulation of the cerebrovascular tone was studied in vitro. Selective inhibition of nNOS by 7-nitro indazole monosodium salt (7-NINA) failed to alter the resting tension and the relaxant effect of bradykinin in isolated rat middle cerebral arteries. These results indicate that 1./ 7-NINA is selective for nNOS and 2./ cerebrovascular nNOS is involved neither in the resting NO production nor in the mediation of the relaxant effect of bradykinin. Therefore, nNOS-derived NO that contributes to the maintenance of the resting cerebral blood flow in vivo appears to be released from neurons and/or glial cells.

All components of an intracerebral kallikrein-kinin system have been described. Thus, bradykinin (BK) acting from the parenchymal side as well as from the blood side may influence cerebral microcirculation. BK is a potent dilator of extra- and intraparenchymal cerebral arteries when acting from the perivascular side. The vasomotor effect of BK is mediated by B2 receptors which appear to be located at the abluminal membrane of the endothelial cell. Signal transmission from the endothelial to the smooth muscle cell is mediated by NO, prostanoids, free radicals or H2O2 depending on the animal species and on the location of the artery. Selective opening of the blood-brain barrier for small tracers (Na+-fluorescein: MW, 376) has been found in cats during cortical superfusion or intraarterial application of BK. This leakage is mediated by B2 receptors located at the luminal and abluminal membrane of the endothelial cells and probably mediated by an opening of tight junctions. Formation of brain edema has been found after ventriculo-cisternal perfusion or interstitial infusion of BK. This can be explained by increase of vascular permeability and cerebral blood flow due to arterial dilatation thus enhancing driving forces for the extravasation. An increase of the BK concentration in the interstitial space of the brain up to concentrations which induce extravasation, dilatation and edema formation has been found under several pathological conditions. Thus, BK may be involved in edema and necrosis formation after cold lesion, concussive brain injury, traumatic spinal cord and ischemic brain injury.

The aim of this study was to investigate the involvement of endothelins (ET) in brain injury. The effect of ET was studied in the isolated basilar artery (BA) taken from control, sham-operated, and cold-lesioned rats. Cold lesion was induced by application of a precooled (-78 degrees C) copper cylinder (outer diameter 5 mm) for 60 seconds to the intact dura over the parietal cortex. After precontraction with prostaglandin (PG) F2alpha, ET-3 (10(-10) to 10(-8) mol/L) dilated BA with a pD2 (negative log of the half-maximal concentration) of 9.06+/-0.031 (mean +/- SD) and a maximal effect (Emax) of 1.64+/-1.0 mN at 3 x 10(-9) mol/L in sham-operated animals. This dilation was reduced 24 and 48 hours after cold lesion by 33% and 73%, respectively, at 3 x 10(-9) mol/L. The effects of acetylcholine (10(-8) to 10(-4) mol/L) and sodium nitroprusside (10(-3) mol/L) were unaltered. Activation of the ETB receptor in thoracic aorta by the specific agonist IRL 1620 also resulted in a reduced dilation (51% by 48 hours after cold lesion). Reverse transcriptase-polymerase chain reaction of the BA showed unaltered expression of mRNA for the ETB receptor after cold lesion whereas ETB immunoreactivity in BA and in its intraparenchymal arteries was reduced at 24 and 48 hours. In contrast to the reduction of ET-3-induced dilation, the constrictor effects of ET-1 and ET-3 were retained after cold lesion. Endothelin-1 (10(-12) to 10(-6) mol/L) dose-dependently contracted segments of untreated control BA segments under resting conditions with a pD2 of 8.03+/-0.22 and an Emax of 6.35+/-0.70 mN. Further evidence that the constrictor ability of BA was not influenced by cold lesion is given by the unaltered response to 124 mmol/L K+ and 10(-6) mol/L serotonin. We conclude that the ETB receptor of BA after cold lesion is downregulated specifically, apparently at the posttranscriptional level. Because the ETB-mediated dilation in thoracic aorta was also reduced, downregulation of the ETB receptor apparently is not restricted to cerebral arteries. The nitric oxide-cyclic guanosine monophosphate system in BA is, however, intact.

Endothelin (ET) vasoconstricts cerebral vessels potently, an effect mediated by ET(A) receptors on the smooth muscle, although the subsequent signaling cascade is unclear. We tested whether the action of ET-1 is mediated by the phospholipase C (PLC)/protein kinase C (PKC) cascade. Isometric force was measured in vitro in ring segments of rat basilar (BA) and middle cerebral (MCA) arteries and expressed as a percentage of the contraction to 124 mM K+. Concentration-effect curves for the constrictor effect of ET-1 (1 pM = 0.3 microM) in control segments or after 25 minutes preincubation with an inhibitor of PLC (neomycin 100 microM) or PKC (H7 10 microM) were constructed under resting tone. In untreated BA, 100 nM ET-1 induced a contraction of 119 +/- 5.3% that fell significantly to 97 +/- 2.8% and 98 +/- 6.7% after neomycin or H7 pretreatment, respectively. In MCA, 100 nM ET-1 induced a contraction of 105 +/- 3.2% that fell significantly to 93 +/- 6.3% and 64 +/- 8.1% after neomycin or H7, respectively. There was no significant shift of the ET-1 EC50 after PKC inhibition in either vessel or PLC inhibition in BA. In summary, the amplitude of ET-1-induced contraction in cerebral vessels is reduced significantly, whereas the sensitivity to the agonist is unchanged, after blocking PLC with neomycin or PKC with H7. This indicates noncompetitive inhibition. ET-1-induced contraction in cerebral vessels thus depends on activation of the PLC/PKC cascade.

Vasodilation elicited by bradykinin (BK) or acetylcholine (Ach) (10 nM-10 microM) in isolated rat cerebral arteries was studied under control conditions, after sham treatment, and after cold lesion (placing a cooled metal probe on the exposed dura) of the cortex. After 24 or 48 hours, isometric force was measured in ring segments of basilar (BA) and middle cerebral arteries (MCA). Concentration-effect curves were constructed after precontraction with 100 microM uridine triphosphate (MCA) or 1 microM serotonin (BA). In MCA and BA, BK elicited similar relative relaxations with maxima of 40.9 +/- 1.5% and 40.7 +/- 3.1%, respectively, at 1 microM. Ach-induced relaxation in BA was much stronger with 82.0 +/- 5.8% at 1 microM. MCA did not relax consistently to Ach. Relaxation to BK in MCA segments was not different between sham-treated and untreated animals. After cold lesion, the dilation to BK (1 microM) was significantly reduced at 24 hours from 0.7 +/- 0.06 to 0.4 +/- 0.06 mN. At 48 hours, this decrease was partly reversed (to 0.5 +/- 0.07 mN). In BA, there was no difference in Ach-induced relaxation between cold-lesioned or sham-treated animals. In summary, the nitric oxide (NO)-mediated response to BK in MCA is attenuated 24 hours after cold lesion. This damage to the BK/NO system is partly reversed 48 hours after the lesion.

This study investigated the role of thromboxane A2 (TXA2) and neuronal nitric oxide (NO) synthase (nNOS)-derived NO in the maintenance of resting cerebrovascular tone. Rat basilar artery (BA) segments were mounted in myographs to study their isometric tension development. 7-Nitro indazole monosodium salt (7-NINA), a specific inhibitor of nNOS, had no significant effect on the resting tone, whereas the general NOS blocker N(G)-nitro-L-arginine (L-NA) induced strong contraction. The thromboxane (TP) receptor antagonist ICI 192605 induced weak vasodilation, and this effect was significantly enhanced after precontraction of the vessels with uridine-5'-triphosphate (UTP). Incubation of BA segments with ICI 192605 attenuated the contractile effect of UTP. These data indicate that nNOS is not involved in resting cerebrovascular NO production and that basal TXA2 release induces a weak contractile tone in the rat BA. Activation of P2U receptors by UTP appears to stimulate TXA2 release in these vessels.

Inhibition of nitric oxide (NO) synthesis induces vasoconstriction and reduction of the blood flow in the brain, indicating that basal release of NO provides a resting vasorelaxant tone in the cerebral circulation. In the present study, the contractile effect of the NO synthase blocker NG-nitro-L-arginine (100 mumol/L) in isolated rat middle cerebral arteries was attenuated markedly in the presence of the cyclooxygenase inhibitor indomethacin (5 mumol/L), the thromboxane A2 synthase inhibitor ridogrel (10 mumol/L), or the thromboxane receptor antagonist ICI 192605 (100 mumol/L). These results indicate that removal of the endogenous NO stimulates the release of thromboxane A2 in cerebral vessels and basal NO production regulates the resting cerebrovascular tone, at least in part, by suppressing thromboxane A2.